Design of Fire Risk Estimation Method Based on Facility Data for Thermal Power Plants

Area of Science:

• Engineering
• Data Science
• Fire Safety

Background:

• Thermal power plants present complex fire risks due to diverse equipment and operational states.
• Existing fire protection systems require robust methods for identifying and classifying potential fire hazards.
• Data-driven approaches can enhance the accuracy and efficiency of fire risk assessment in industrial facilities.

Purpose of the Study:

• To propose a novel data classification and analysis method for estimating fire risk in thermal power plants.
• To develop a system that identifies, classifies, and integrates fire risks into existing thermal power plant safety protocols.
• To provide a framework applicable to different thermal power plant sizes and configurations.

Main Methods:

• Facilities were categorized into three states: Steady, Transient, and Anomaly, based on operational conditions.
• Thermal power plants were zoned into turbine, boiler, and indoor coal shed areas, with further subdivision into specific equipment.
• Fire-related tags from Supervisory Control and Data Acquisition (SCADA) data were utilized, focusing on pool, 3D, and jet fire scenarios.

Main Results:

• A method was developed to classify fire risks across different zones and equipment within thermal power plants.
• Three fire hazard levels were organized for each zone, based on analysis of historical fire and explosion scenarios.
• Experimental analysis was performed on data from 500 MW and 100 MW thermal power plants.

Conclusions:

• The proposed data classification and analysis method effectively estimates fire risk in thermal power plants.
• This approach offers valuable insights for data analysts, both with and without domain expertise in power plant fires.
• The method provides a foundation for enhancing fire protection systems in thermal power plants.